Abstract
This study aimed to understand how the COVID-19 pandemic alters day-to-day relational experiences and how daily relational experiences shape outlook on the pandemic. Data were collected from university students in the U.S. using smartphone-based experience sampling and nightly diary surveys over a 10-day period beginning in April 2020. To address the first objective, we examined how pandemic-related anxiety and depressive symptoms manifested in three aspects of daily relational life: (a) perceptions of loneliness, (b) relational difficulty, and (c) communication quality. To address the second objective, we investigated how daily relational experiences were linked to their end-of-day outlook of the pandemic in terms of pessimism versus optimism regarding COVID-19 and general sense of hope. Over the 10-day study period, COVID-19-related anxiety was found to be uniquely associated with greater loneliness, desire for interaction when alone, interpersonal conflict, and distraction during communication episodes. In addition, results showed that relational experiences throughout the day, including loneliness, challenges maintaining relationships, and communication quality, were linked to end-of-day COVID-19 optimism and state hope. Overall, this study identifies potential ways in which young adults’ relational lives were altered during the first wave of the COVID-19 pandemic in the U.S. Results also suggest how relational experiences factor into daily social constructions of the pandemic.
Keywords
In early 2020, the COVID-19 (SARS-CoV-2 coronavirus) outbreak was declared a global pandemic by the World Health Organization. Since that time, the virus has caused over three million deaths worldwide, devastated large swaths of the global economy, and spawned collective trauma as people restructure their lives in efforts to slow the virus’s spread (Watson et al., 2020; World Health Organization, 2021). The pandemic has also fundamentally reshaped relational life.
Two key questions regarding COVID-19 and relational life guide the current study. First, we sought to understand how the pandemic alters day-to-day relational experiences. Second, we aimed to identify how daily relational experiences shape outlook on the pandemic. To address the first question, we examine how pandemic-related anxiety and depressive symptoms manifested in three key aspects of daily relational life: (a) perceptions of loneliness, (b) relational difficulty, and (c) interpersonal communication quality. To address the second question, we investigated how daily relational experiences in those same three areas (i.e., loneliness, relational difficulty, and communication quality) were linked to end-of-day outlook on the pandemic in terms of pessimism versus optimism regarding COVID-19 and sense of hope.
Perception of the pandemic is influenced by myriad factors, including personal experiences with COVID-19 infection, financial (in)security, and the nature of news and government reports about infection rates (Van Bavel et al., 2020; Walsh, 2020). In addition to these factors, day-to-day relational experiences likely also influence how people view the pandemic. Whereas days filled with positive and rewarding face-to-face or mediated communication with loved ones might lead individuals to feel better able to endure the challenges and restrictions imposed by the pandemic, the opposite might be true of individuals whose days are filled with difficult, unfulfilling, or insufficient communication. In short, when daily relational life is challenging during a pandemic, it might provoke a sense of hopelessness about the pandemic itself and life in general.
To capture daily relational experiences during the early months of the pandemic in the U.S., data for this study were collected from a sample of undergraduate students using the experience sampling methodology (ESM; Hektner et al., 2007). Participants were surveyed at random moments each day over a 10-day period via a smartphone app on their personal device. They reported their current thoughts, emotions, and behavior across several thousand moments of interaction and alone time. Participants also completed a presurvey that provided data on their mental health as well as nightly diary surveys that solicited their end-of-day perceptions and cognitions. These varied survey approaches provided insight into participants’ general perceptions of the pandemic as well as their moment-to-moment experiences of it. ESM data is especially useful because, in addition to providing data at both within-person and between-person levels (Bolger & Laurenceau, 2013), it enables us to examine nuanced features of participants’ relational experiences (e.g., experiences alone vs. in the company of others, how experiences during the day are linked to end-of-day perceptions). The data for this study were collected in the U.S., between April and June of 2020. This was a period of rising COVID-19 infections and deaths. COVID-19 deaths in the U.S. surpassed 50,000 at the start of April and 100,00 by the beginning of June (Taylor, 2021).
The COVID-19 pandemic and college student populations
College students, in general, experience high rates of mental health challenges, and research indicates that mental health diagnoses, such anxiety and depression, have steadily increased over time (Lederer et al., 2021; Oswalt et al., 2020). Due to such factors as increased social isolation, altered living arrangements, financial hardships, and heightened health concerns for themselves and loved ones, the COVID-19 pandemic appears to have only worsened the situation (Aristovnik et al., 2020; Browning et al., 2021; Wang et al., 2020). As Lederer and colleagues (2021) put it, the COVID-19 pandemic “has exacerbated college students’ known mental health risk factors and other health conditions” (p. 15).
Even for students who have access to safe living environments (e.g., with family members), taking classes remotely presented new challenges, especially for students lacking sufficient internet access or adequate access to work space (Lederer et al., 20201). Many students also took on additional caregiving roles in their families (for both older and younger individuals) during the pandemic, while often also confronting financial burdens (Aristovnik et al., 2020; Lederer et al., 2021). Consider, too, that some students were forced to return to home environments in which family members might not be fully supportive of their identities (Lederer et al., 2021). Overall, there is great heterogeneity in people’s experiences during the pandemic. It seems clear, however, that college students have experienced multiple challenges that are risk factors for deteriorated mental and physical health.
COVID-19-related anxiety and depressive symptoms and daily experiences
Our first aim in this study is to understand how the pandemic can intrude upon daily relational experiences. Research shows that there have been widespread increases in reported depression and anxiety since the outbreak of COVID-19 (Bäuerle et al., 2020; Salari et al., 2020). Nearly a quarter of participants from an early metanalysis on COVID-19 reported symptoms of anxiety and depression related to the pandemic, making COVID-19-induced anxiety and depression quite common (Rajkumar, 2020).
Depression and anxiety, while often comorbid, are distinguishable (Palgi et al., 2020). Although both are described as forms of maladaptive cognitive processing, depression concentrates on an “exaggerated appraisal of loss” and is characterized by negativity toward self and the world (Beck & Clark, 1988, p. 24). Anxious individuals, in contrast, overestimate threats associated with a situation or event. Beck and Rush (1988) suggested that anxiety negates an individual’s potential to withdraw and conserve energy when needed, often leading to a depletion of resources.
In the current study, we examine how anxiety and depressive symptoms stemming from the pandemic are associated with three aspects of relational life: loneliness, relational difficulty, and communication quality. We focus on these aspects because they are core to the study of social and personal relationships (Finkel et al., 2017) and are key socio-communicative components of well-being (Merolla et al., 2021b). In efforts to comprehensively assess loneliness, relational difficulty, and communication quality, we measure multiple components of each. For loneliness, we examine (a) generalized feelings of loneliness across random moments of the day and (b) desires for social interaction during moments of alone time. Desiring interaction when one is alone is a unique indicator of loneliness that taps into unsatiated needs for belongingness (Hall & Davis, 2017). For relational difficulty, we assess (a) the amount of interpersonal conflict that people experience over the course of a day along with (b) perceived challenges in successfully maintaining relationships throughout the day. When people experience high levels of conflict and struggle to satisfactorily maintain relationships over the course of a day, it creates social conditions for diminished well-being (Darbonne et al., 2013; Ryff, 1989). Finally, to assess communication quality, we measure (a) how much satisfaction participants drew from each randomly-sampled interaction they reported and (b) how distracted they were during the interaction. Distraction can reflect low momentary communication competence and reflects an interaction style negatively linked to well-being (Merolla et al., 2019).
We propose that each one of the three aspects of relational life will be linked to COVID-19-related anxiety and depressive symptoms. The cognitive-behavioral model of health anxiety (Taylor & Asmundson, 2004), for example, suggests that the pandemic could act as a triggering event that sparks health anxiety throughout the day, resulting in negative emotional and relational experiences. Although anxiety can be quite functional and help people avoid risky behavior, particularly in the context of a pandemic, the functional benefits of anxiety can also come with social and relational costs (Baumeister & Tice, 1990). Escalated levels of COVID-19-related health anxiety, for example, have been found to be significantly associated with feelings of isolation and thwarted needs for belongingness (González-Sanguino et al., 2020; Palgi et al., 2020). Recent work also suggests that heightened focus on COVID-19-related health anxiety can increase the probability of destructive and distracted communication (Pietromonaco & Overall, 2020). As Pietromonaco and Overall (2020) proposed, anxiety and stress attributable to the COVID-19 pandemic can “can create a context in which it is more difficult for partners to be responsive to each other because they are distracted, fatigued, or overwhelmed” (p. 4). In sum, anxiety about COVID-19 can undermine daily relational experience by making individuals more likely to feel alone and unsupported, less capable of engaging in constructive relational processes, and more likely to be cognitively preoccupied by fear and worry during interpersonal encounters.
Research indicates that depressive symptoms stemming from COVID-19 will evidence similar effects on relational experiences. This is not only because of the aforementioned comorbidity of anxiety and depressive symptoms, but also because of the specific ways in which depressive symptoms are theorized to emerge interpersonally for many individuals (Coyne, 1976; Nolen-Hoeksema et al., 2008; Weissman et al., 2000). Studies have found that an increase in depressive symptoms tends to be associated with spending less time in social interaction (Libet & Lewinsohn, 1973), participating in fewer interactions throughout the day, and engaging in interaction with fewer people at a time (e.g., more dyadic conversations compared to group conversations; Elmer & Stadtfeld, 2020). These factors can exacerbate existing feelings of social exclusion (Baumeister & Tice, 1990).
Scholarship on depression in close relationships also suggests that depression is typically negatively associated with relational quality (Nolen-Hoeksema et al., 2008; Segrin, 2000). Individuals with depression have been found to interact with people in ways that provoke negative reactions from others (e.g., dismissal, rejection, exasperation; Coyne, 1976; Nolen-Hoeksema et al., 2008). Depression, for instance, predicts harmful communication patterns with a romantic partner, such as hostile conflict (Byrne et al., 2004; Sherabi et al., 2016). Empirical work has also demonstrated a link between depression and social skill deficiency, negative attributions of others’ communication, disparaging assessment of one’s own communication, and tendencies to negatively perseverate about past interactions (Nolen-Hoeksema et al., 2008; Steger & Kashdan, 2009).
Based on the work reviewed, we predicted that COVID-19-related anxiety and depressive symptoms will be positively associated with indicators of loneliness and relational difficulty and negatively associated with indicators of communication quality. A conceptual model of the following hypotheses is presented in Figure 1.
Depiction of the proposed associations (H1–H6) and the data sources for study variables.
Daily relational experiences and end-of-day perceptions
Next, we endeavored to understand how daily relational experiences are linked to people’s end-of-day perceptions. At the time of data collection, the prospects of a vaccine were unclear, most schools and business were closed in a first wave of lockdowns, and stay-at-home orders were in place throughout much of the U.S. Uncertainty, in sum, was quite high, and it was easy to feel a sense of hopelessness about whether or not and how the pandemic could be successfully managed (Koffman et al., 2020). We therefore chose to examine the daily factors, specifically within the domain of personal relationships, associated with people’s end-of-day optimism about the pandemic and general sense of hope. Literature on resilience and thriving contends that personal relationships are key sources of meaning, inspiration, and coping in challenging times (Afifi, 2018; Buzzanell, 2010). In difficult times, people turn to their personal relationships for tangible and emotional resources (Hobfoll, 1998) that can help them sustain a sense of optimism and hope to persevere (Beck, 2016). Among university students, in particular, recent research suggests that although their levels of stress have increased during the pandemic, their utilization of on- and off-campus mental health services has not kept pace (Lee et al., 2021). It is plausible, then, that students have increasingly turned to other sources, such as friends and family, for support in managing their increased stress levels.
It also appears to be the case that hope and optimism can be dashed during difficult times, such as the current pandemic, when relational experiences are tumultuous (Pietromonaco & Overall, 2020; Rodriguez-Hanley & Snyder, 2000). Optimism and hope, in short, are sensitive to the nature of people’s relational experiences (Snyder & Feldman, 2000). In following sections, we therefore consider how daily relational experiences might be related to end-of-day COVID-19 optimism and general state hope. A conceptual model of the predictions discussed below is presented in Figure 2.
Depiction of the proposed associations (H7–H12) and the data sources for study variables.
COVID-19 optimism
Optimism is a general expectation of positive experiences in the future (Carver et al., 2010). Maintaining optimism about the COVID-19 pandemic can be critical for sustaining well-being (Biber et al., 2020; Jovancević & Milićević, 2020). An optimistic outlook, unlike an outlook marked by fatalism and despair, can motivate people to adhere to positive health related-behaviors (Carver et al., 2010). During the pandemic, an optimistic outlook, whereby one feels as if pandemic-based restrictions are manageable because a more positive future is on the horizon, might translate to better adherence to public health guidelines and mandates (Millstein et al., 2016).
To be clear, we are not arguing here about unrealistic optimism, a cognitive distortion whereby individuals view themselves as less vulnerable than others in a manner than can lead to problematic health behaviors and non-adherence to public health recommendations (Dolinski et al., 2020). Rather, we are focusing on optimism-as-perseverance in the face of an uncertain and challenging situation (Mead et al., 2020). This is akin to what Frankl (1984) termed “tragic optimism,” or the ability to maintain a sense of meaning and belief in a better future in spite of suffering and tragedy. Sustaining tragic optimism was argued by Mead et al. (2020) to be an essential feature of coping and well-being during the COVID-19 pandemic.
Hope
In addition to optimism about the pandemic itself, we are also interested in people’s general sense of state hope at the end of each day. Although optimism and hope share conceptual similarities, they are argued to be unique constructs (Alarcon et al., 2013). Whereas optimism is conceptualized as a generalized sense of positive expectations about the future (Carver & Scheier, 1998), hope is typically conceptualized as a goal-based construct rooted in individuals’ ability to devise and pursue goals they believe are important (Snyder, 2002). A goal-centric view of hope is most often associated with the influential work of Snyder and colleagues’ hope theory (Snyder & Feldman, 2000). Hope theory proposes that there are two key components of the hopeful mindset: agency and pathways. Agency hope is the sense of motivation necessary to pursue routes toward goals (despite any roadblocks that might arise), whereas pathways hope consists of the specific routes toward goals that individuals develop (Snyder, 2002). Hope theory conceptualizes agency and pathways as critical components of resilience in the face of adversity (Snyder, 2002). Consistent with this view, research within a hope theory framework has shown that high agency and pathways hope are linked to positive outcomes in difficult situations, ranging from coping with serious injuries (Barnum et al., 1998) to recovery from natural disasters (Hackbarth et al., 2012).
Consistent with hope theory, recent research on the COVID-19 pandemic has shown that that agency and pathways are positively associated with successful coping with stressors (Yildirim & Arslan, 2020). Research, however, has not examined the specific factors that support hope during the pandemic. Thus, our interest in this study is to determine how end-of-day state hope (i.e., agency and pathways thinking) are associated with relational experiences over the course of the day. Hope theory proposes that hope necessitates the combination of agency and pathways, such that individuals benefit most from high motivation to pursue multiple routes toward goals. In our hypotheses, we predict that the direction of effects will be the same for agency and pathways. Testing agency and pathways separately (as opposed to merged) is beneficial, however, because it enables us to determine if the study variables are associated with only agency or pathways—such insight is useful for continued work on hope theory. For end-of-day COVID-19 optimism and state hope (agency and pathways), we offer the following hypotheses (see also Figure 2).
Method
Participants and procedures
Participants were 120 upper-division students at a large university in California (M age = 21, SD age = 1.11). Participants were free to report multiple races and ethnicities; approximately 46% identified as Asian, 46% White, 18% Latina/o/x, 5% Black/African American, 5% Native American, and 2% Native Hawaiian or Other Pacific Islander. Eighty-one percent of participants identified as female and 19% identified as male. Following email announcements, preliminary phone calls were scheduled with interested individuals. During the call, a research assistant reviewed the study requirements and incentive. For full participation, participants received a $40 Amazon e-gift card. Participants were then asked to download a smartphone app from the Google Play or Apple stores that would serve as the primary data collection apparatus for the study.
The survey app uses the native notification systems on Android and Apple devices to signal participants when it is time to complete a survey. The app was setup to guide participants through the study. On day 1 (Sunday morning), participants received the presurvey, which solicited demographic information and responses to various self-report measures. On day 2, participants began their 10-day experience sampling/diary period. Six experience sampling surveys were sent to them at random times within specified time intervals between 9:00 AM and 7:00 PM. Each night, participants also received a seventh survey—a diary survey—that asked them to report on the day’s events. Unrelated to the current study, beginning on day 4 of the experience sampling/diary period, some of the participants received messages in the app in addition to the surveys. These messages were part of a pilot study on positive communication. Our tests indicated that receiving or not receiving these messages was not significantly related to any of the study variables. All models reported in this manuscript also control for any potential influence of condition (as will be seen, no significant relationships emerged based on condition). This is the first manuscript being written from this dataset.
Measures
Three survey sources (i.e., presurvey, experience sampling surveys, and nightly surveys) were used given that some items were best answered in one of the three formats. Given that measures were completed on participants’ phones, and in keeping with best practices of ESM (Hektner et al., 2007), single or two-item measures were employed throughout.
Presurvey
COVID-19-related anxiety and depressive symptoms
These measures were adapted from the brief Patient Health Questionnaire (Kroenke et al., 2009), which contains two items assessing anxiety (i.e., Feeling nervous, anxious, or on edge; Not being able to stop or control worrying”) and two assessing depression (i.e., Feeling down, depressed, or hopeless; Little interest or pleasure in doing things). Participants were presented the following instructions: “Over the past 10 days, due specifically to the COVID-19 pandemic, how often have you been bothered by the following?” Responses options were on a five-point scale ranging from Not at all (1) to Everyday (5). Consistent with past research (Kroenke et al., 2009), anxiety (M = 2.78, SD = 1.14, α = .85) and depressive symptoms (M = 2.49, SD = 1.09, α = .79) were moderately correlated (r = .57, p < .001).
Experience sampling survey
Momentary loneliness
Loneliness was measured at each survey instance. Participants were asked, “At this moment, how lonely do you feel?”. Responses (M = 2.50, SD = 1.53) were on a seven-point scale, ranging from Not lonely at all (1) to Very lonely (7).
Desire for social interaction when alone
If participants indicated that they were currently alone, they were presented with the following question, “Do you wish you were engaged in interaction with someone right now or are you content being alone?” Responses (M = 2.63, SD = 1.88) ranged from Content being alone (1) to Wish I were interacting with someone (7).
Distraction during social interactions
If participants indicated that they were in an interaction, or had one within the past 10 minutes, they were asked to report how “distracted” they felt during the interaction. Response options ranged from Not at all (1) to Extremely (7) (M = 2.17, SD = 1.11).
Perceived satisfaction with social interactions
For any interaction reported on, participants were asked to rate the degree to which they were “satisfied with the interaction.” Responses (M = 3.56, SD = 1.02) ranged from Not at all (1) to Extremely (7).
Nightly diary survey
Daily challenges in maintaining relationships
Each night, participants were asked to report on their relational experiences over the course of the day, including relational difficulty. This item, based on Ryff’s (1989) work on well-being, stated, “Maintaining relationships was difficult and frustrating today.” Participants responded on a scale ranging from Strongly disagree (1) and Strongly agree (7) (M = 2.98, SD = 1.62).
Daily interpersonal conflict
Participants also reported on their conflict experiences based on this question: “Did you have any conflicts or arguments with people today?” Responses (M = 1.46, SD = .68) options ranged from None (1) to A great deal (7).
End-of-day optimism regarding the pandemic
The nightly survey also asked participants to report their end-of-day outlook on the pandemic: “When you thought about the COVID-19 pandemic today, did you feel optimistic or pessimistic about the situation?” The responses (M = 3.49, SD = 1.39) were anchored by Very pessimistic (1) to Very optimistic (7).
End-of-day state hope
Based on Snyder’s (2002) measures of hope, agency hope (M = 4.08, SD = 1.59) was measured with the item, “Today, I energetically pursued my goals” and pathways hope (M = 4.07, SD = 1.33) was measured with the item, “Today, I felt like there were lots of ways around any problems I faced,” These two items were used in previous diary research (Merolla et al., 2021a). Participants responded with a seven-point scale ranging from Strongly disagree (1) to Strongly agree (7).
Results
Descriptive statistics
For the experience sampling surveys, participants completed 6,108 survey instances over the 10 days, putting the experience sampling survey response rate at approximately 85%, which is considered very good. Participants reported on a social interaction in approximately 55% of these moments, with the remaining 45% being moments alone. Of the moments of interaction, 30% of them were face-to-face, 6.3% were text-based (e.g., text messages/SMS), 6.1% video calls (e.g., Facetime), and 2.1% were voice-only calls. For the nightly diary survey, which was sent each night over the 10 days, the response rate was approximately 93% (1,112 out of 1,200). The presurvey scales of COVID-related anxiety and depressive symptoms were completed by 115 of the 120 participants (96%).
Analytic plan
The data for this study are nested, such that survey responses from the experience sampling and nightly diaries are embedded within participants over time. Thus, the data were analyzed using multilevel modeling (with random intercepts). Models were run in R with the lme4 package (version 1.1-21; Bates et al., 2015). Given the multilevel structure of the data, variables collected through the experience sampling or the nightly surveys needed to be separated into within-person and between-person versions (Bolger & Laurenceau, 2013). Within-person versions of the variables are considered level 1 data (i.e., the moment/day level). Data at level 1 capture within-person change, such that each score for a variable at this level represents a deviation from the participant’s average or “typical” score on that variable. The between-person versions of variables are considered level 2 data (i.e., the person level). Data at level 2 reflect overall mean-level differences between participants on given variables. Anytime a variable collected through experience sampling or the nightly surveys serves as an independent variable in the analysis (i.e., models for H7–H12), the within-person and between-person versions of the variables were entered into the models. In contrast, when the experience sampling or nightly survey variables serve as dependent variables in the models (i.e., H1–H6), the raw score of these variables was utilized. If models contain within-person (i.e., level 1) independent variables (see Tables 4–6), the within-person estimates will be of primary interest (Bolger & Laurenceau, 2013). In models where variables from the pre-survey were entered as independent variables (i.e., COVID-19-related anxiety and depressive symptoms; see Tables 1– 3), they were grand mean centered. Throughout the analyses, we control for time in two ways. First, we controlled for survey number to control for any effects attributable to time spent in the study. Second, we controlled for the date in which participants began the study.
COVID-19-related anxiety and depressive symptoms predicting loneliness indicators (H1 and H2).
Note. ***p < .001, **p < .01, *p < .05.
COVID-19-related anxiety and depressive symptoms predicting relational difficulty indicators (H3 and H4).
Note. ***p < .001, **p < .01, *p < .05.
COVID-19-related anxiety and depressive symptoms predicting communication quality indicators (H5 and H6).
Note. ***p < .001, **p < .01, *p < .05.
Hypothesis tests
COVID-19-related anxiety and depressive symptoms
The first six hypotheses concerned associations between pandemic anxiety and depressive symptoms and the relational variables (see Figure 1). As noted earlier, many scholars argue that anxiety and depression are often comorbid. Given this, along with the fact that we adapted the existing anxiety and depression measures into the COVID-19 context, we tested H1–H6 in models where anxiety and depressive symptoms were isolated and in models in which they were jointly analyzed. We reasoned that this approach provides future researchers maximum information regarding correlates of pandemic-related anxiety and depression. We will contextualize the results based on these varying models (see Tables 1–3).
Loneliness and desire for interaction when alone (H1 & H2)
As shown in Table 1, COVID-19-related anxiety and depressive symptoms were significantly and positively associated with loneliness and desire for interaction when entered in the models separately. In the combined model, however, only anxiety remained significantly and positively associated with loneliness and desire for interaction. The individual models supported H1 and H2. Based on the combined model, though, only H1 was supported.
Conflict and maintenance challenges (H3 & H4)
As shown in Table 2, in the anxiety-only model, anxiety was positively associated with both daily conflict and relational challenges; this pattern supports H3 in full. In the combined model, however, anxiety was only positively associated with daily conflict, which lends support only for H3a. In the depressive symptoms-only model, depressive symptoms was positively associated with relational challenges, but not daily conflict, thus supporting H4b, but not H4a. In the combined model, however, depressive symptoms became a non-significant predictor of daily conflict and relational challenges, which fails to support all components of H4.
Communication satisfaction and distraction (H5 & H6)
As shown in Table 3, in the anxiety-only model, anxiety was significantly and positively associated with distraction during social interactions, but non-significantly associated with communication quality. This same pattern was found it the combined model, thus supporting H5b, but not H5a. In the depressive symptoms-only model, depressive symptoms was significantly and positively associated with distraction during interaction, but non-significantly associated with communication quality, which is consistent with H6b, but not H6a. In the combined model, however, depressive symptoms became non-significantly associated with distraction and communication quality, a pattern that failed to support all aspects of H6. 1
Outlook on the pandemic: Optimism and agency and pathways hope
Loneliness and desire for interaction when alone (H7 & H8)
In order to test H7 and H8 through multilevel modeling (with day-level dependent variables), the experience sampling moments for independent variables of loneliness and desire for social interaction were calculated at the day-level (i.e., averaging scores from the moment level to the day level). As shown in Table 4, results supported H7a and H8a, as within-person loneliness was significantly and negatively associated with end-of-day pandemic optimism, pathways hope, and agency hope. Loneliness at the between-person (or mean) level was also significantly and negatively associated with pathways and agency hope, though that result is of less interest. Desire for interaction when alone (at the within-person level), was, however, significantly and negatively associated with pathways hope; this partially supports H8b. Desire for interaction was not significantly related to optimism, which fails to support H7b. Interestingly, between-person desire for interaction was positively associated with end-of-day agency, which indicates that participants with higher overall mean levels of desire for interaction also reported higher overall mean levels of agency over the 10-day period. This result demonstrates the value of conducting analyses at both within- and between-persons levels of analysis, as the two levels can evidence results of differing directionality (Bolger & Laurenceau, 2013). The within-person estimates, however, are most relevant to the hypothesis tests.
Loneliness and desire for interaction (when alone) throughout the day as predictors of end-of-day COVID-19 optimism and agency and pathways hope (H7 and H8).
Note. ***p < .001, **p < .01, *p < .05.
Maintenance challenges and conflict (H9 & H10)
Since relational maintenance challenges and daily conflict were from the nightly survey, they were already at the day-level; thus, no averaging was required. As shown in Table 5, H9a and H9b were supported, as both relational challenges and daily conflict—at the within-person level—were significantly and negatively associated with end-of-day optimism regarding the pandemic. H10a was supported, as within-person maintenance challenges was significantly and negatively associated with end-of-day agency and pathways hope. H10b was partially supported, as within-person conflict was significantly and negatively associated with agency hope, but not pathways hope. Of note, there were also between-person negative associations between maintenance challenges and optimism, agency hope, and pathways hope.
Daily relational challenges and conflict as predictors of end-of-day COVID-19 optimism and agency and pathways hope (H9 and H10).
Note. ***p < .001, **p < .01, *p < .05.
Distraction and perceived communication satisfaction (H11 & H12)
Prior to analysis, distraction and communication quality were computed at the day-level. H11a and H11b, regarding COVID-19 optimism, were supported. As shown in Table 6, within-person distraction was negatively associated, and within-person communication satisfaction was positively associated, with end-of-day optimism regarding the pandemic. H12a, which hypothesized that within-person distraction would be negatively associated with agency and pathways hope, was not supported. H12b, in contrast, was supported, as within-person communication satisfaction was positively associated with end-of-day agency and pathways hope. There were also several significant between-person associations (in the expected directions) in the models.
Distraction during interaction and perceived interaction quality throughout the day as predictors of end-of-day COVID-19 optimism and agency and pathways hope (H11 and H12).
Note. ***p < .001, **p < .01, *p < .05.
Discussion
Over the 10-day study period, COVID-19-related anxiety was found to be uniquely associated with greater loneliness, desire for interaction when alone, interpersonal conflict, and distraction during communication episodes. In addition, results showed that relational experiences throughout the day were linked to end-of-day COVID-19 optimism and general state hope (agency and pathways). Overall, this study suggests that for the young adult participants in this study, relational experiences are important components of the COVID-19 pandemic; not only does the pandemic appear to intrude upon daily relational experiences, but daily relational experiences also appear to shape perspective on the pandemic itself.
The findings regarding COVID-19 anxiety suggest that increased fear, worry, and rumination about the pandemic situation tend to adversely affect relationships. Taylor et al. (2020) reported that individuals with health anxiety related to COVID-19 reported higher rates of self-imposed isolation and excessive avoidance of public places above and beyond the recommendation of the CDC. Although such precautions can be warranted and beneficial for keeping people safe from infection, they might also exact an interpersonal toll. Intense preoccupation and perseverative cognition concerning the virus might decrease participation in social interaction while also depleting the energy needed to enact positive and constructive relational communication (Hall & Davis, 2017; Merolla et al., 2019). Participants in this study might have found themselves newly isolated from close friends or separated from family members whom they were trying to protect. Heath anxiety under such conditions can be especially taxing both intra- and interpersonally.
Although both COVID-19-related anxiety and depressive symptoms were significantly linked to the relational experience variables when analyzed in separate models, only anxiety remained significant when depressive symptoms and anxiety were jointly analyzed. Depressive symptoms, in fact, were not significantly related to any of the dependent variables in the combined models. Why might this be? One possibility concerns the overlapping cognitive processes operative within anxiety and depression. Given that both anxiety and depression involve negative rumination, for instance, any influence of depressive symptoms on the relational variables could presumably have been channeled through anxious rumination (Nolen-Hoeksema et al., 2008). Thus, when both variables were entered into the models, only anxiety accounted for unique variance.
Another possibility is that the pandemic situation has altered the typical associations that exist between depressive symptoms and forms of relational communication, such as social support. Leading perspectives on depression, such as response styles theory (Nolen-Hoeksema et al., 2008) and the interpersonal theory of depression (Coyne, 1976), suggest that people with depression often receive relatively low levels of social support. This can occur for several reasons. This can occur, for instance, when family members and friends feel overburdened and become frustrated by the persistent brooding and discussion of negative feelings by depressed individuals (Nolen-Hoeksema et al., 2008). On top of this, university students might live away from family and friends who can provide important forms of emotional support while also lacking access to affordable mental health services (on-campus or otherwise). There is also evidence that college students tend to normalize mental health challenges as a natural part of college-related stress, which can reduce motivation to seek out support services (Oswalt et al., 2020). But the pandemic might have reversed some of these trends. In the context of the pandemic, a period that has proven anything but normal, people might have given increased attention and support to those individuals in their lives who were showing signs of depressive symptoms. The atypical nature of the pandemic might have heightened people’s sense of concern and compassion for friends and family members, especially those showing signs of depression related to COVID-19, thus attenuating some of the negative effects depressive symptoms tend to have on personal and relational experiences.
Thus far we have speculated that pandemic-related anxiety contributed to greater loneliness, more interpersonal conflict, and less engagement during social interactions over the 10-day study period. But this is just one interpretation of the data. Despite the fact that anxiety was measured prior to the experience sampling/nightly diary period, the data do not enable us to make claims about causality. It therefore remains entirely possible that loneliness and relational difficulty contribute to COVID-19 anxiety (see Palgi et al., 2020). Indeed, it could be the case that feelings of disconnection from others and difficult daily interaction makes people feel nervous, worried, and fearful about how they can continue managing what feels like a unceasing pandemic.
This study also found that relational experiences throughout the day were linked to end-of-day COVID-19 optimism and general state hope (in the form of agency and pathways), suggesting relational experiences factor into social construction of the pandemic. Results at the within-person level for optimism indicated that individuals felt less optimistic (more pessimistic) about the pandemic on days when they reported higher-than usual loneliness, distraction during interaction, conflict, and maintenance challenges, and lower-than-usual communication satisfaction with interaction. This suggests that the students in our sample perceived the pandemic situation as more manageable and that the future was brighter on days when their relational lives were more positive. Alternatively stated, difficult relational experiences might have made the pandemic situation feel that much bleaker. Of course, it is also plausible that there could be outside factors (i.e., omitted variables in this study) that simultaneously dampened COVID-19 optimism and negatively influenced relational perceptions. Alternative explanations must be taken seriously. For example, it could be the case that on days when news reports were especially dire, people tended to have difficult relational experiences and lower optimism about the pandemic. Studies that can account for cognitive and communication factors as well as factors from the larger communication ecosystem (e.g., news media) could be especially insightful regarding the multilevel influences on people’s pandemic outlook.
Results also indicated that hope (i.e., agency and pathways) was higher on days when relational experiences were better than usual. Specifically, both agency and pathways hope were higher on days when participants reported lower-than-typical loneliness, maintenance challenges, and higher-than-usual communication satisfaction with interaction. Pathways was also negatively associated with desire for interaction when alone, while agency was also negatively associated with daily conflict. These results can be interpreted within resilience and hope theory frameworks. Contemporary resilience perspectives place relationships, including relational and communication quality, front and center in the promotion of thriving during difficult times (Afifi, 2018; Buzzanell, 2010). Similarly, a core assumption of hope theory, albeit one that has not been thoroughly tested, is that hope is built upon a strong foundation of interpersonal connections and support (Snyder & Feldman, 2000). This study provides unique evidence for the ways in which positive daily relational experiences might support state hope. Thus, these findings contribute to the literature on hope theory, particularly among university students—a population that has long been of central interest to hope theorists (Snyder et al., 2002).
The results of this study, while certainly preliminary pending further research, could have practical applications for clinicians and network members of young adults who are at high risk of mental health challenges. As has been noted, participants who reported higher COVID-19-related anxiety not only felt more alone, but also were having difficulties maintaining relationships and having positive social interactions throughout the day. This suggests that it would be beneficial for people to closely monitor interactions with friends and family members for signs of heightened anxiety, loneliness, and distraction. Based on the current results, this might be especially critical for young people who live alone and/or are geographically-separated from family and friends. If, for example, someone demonstrates signs of increased anxiety during the pandemic and begins calling/texting less frequently than usual or gets easily annoyed during conversations, it might be important to put in extra effort to regularly check in with them about their well-being. Providing information on digital access to mental health professionals (e.g., via online services provided by university health clinics) could be helpful.
Various limitations of this study are also important to discuss. Although a strength of the present study was its ability to investigate moment- and day-level relational experiences via a smartphone app, experience sampling and diary surveys involve notable limitations. Many of the measures used to assess the interactions were single items which restricts the comprehensiveness of measurement of the variables. Also, the surveys, despite being completed in close proximity to experiences, are still self-report and therefore subject to cognitive biases. It should also be kept in mind that the data was collected within the first 2 months of state and local lockdowns in the United States. Patterns of communication, and living situations for that matter, among the participants were still being sorted out. This could therefore have been a period of especially high uncertainty in the pandemic. Results might be different in the later stages of the pandemic and in different locations.
Another key limitation of this study is that it sampled university students. As discussed at the outset, university students are a vulnerable group for mental health challenges. Moreover, they have experienced very real challenges during the pandemic. Still, their situations might be markedly different than the situations of other types of people. It is plausible, for instance, that our participants had stronger financial support networks in place than do other groups. Further, childcare, a stressor for many people during the pandemic, might not have been a major factor for many or most of the current participants. Thus, generalizing the current results beyond a university population from California might not be advisable. In general, we must keep in mind that the pandemic has taxed people in different ways, and as a consequence, people’s experiences, including their relational experiences, could be quite heterogeneous. We were also limited in that we did not have data on other important experiences during the pandemic. Research indicates, for instance, that students of color and people from marginalized communities have experienced intensified hardships during the pandemic, including increased racial discrimination (Browning et al., 2021). Considering that many of the participants in this study identified as people of color, survey questions about perceived discrimination could have providing important insights into their daily experiences.
Despite these limitations, future research can hopefully build upon the current results. The experience sampling methodology, conducted through smartphone apps, can be quite useful during the pandemic and other situations necessitating limited face-to-face communication between research staff and participants. This method enables contactless and remote data collection while also providing rich data on people’s relational experiences. Conducting similar studies with various samples can cast further light on the pandemic as it continues to shape the intra- and interpersonal lives of people around the world.
Footnotes
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a grant from the University of California, Santa Barbara Academic Senate.
Open research statement
As part of IARR’s encouragement of open research practices, the authors have provided the following information: This research was not pre-registered. The data used in the research are available. The data can be obtained by emailing: